FW: a standard to publish data on the Web

[David Blankman]

David Blankman
921 Ortega, NW
Albuquerque, NM 87114-1417
(505) 246-1625 ph (505) 450-6045 mobile

-----Original Message-----
From: Peter McCartney [mailto:peter.mccartney at asu.edu]
Sent: Thursday, February 21, 2002 1:52 PM
To: 'Eda Melendez-Colom'; dman at lternet.edu
Subject: RE: a standard to publish data on the Web

There is no standard - but im sure Microsoft and ESRI would be happy to work
with us to define one!

There are some constraints that we might consider more formally specifying
using EML as a reference. One deals with defining guidlines for logical
entity types.  Currently, EML recognizes one logical entity type - a table,
consisting of rows and columns. We have poposed several others such as
gislayer, grid, view, etc. A complicated table with subtotals, breaks, or
hierarchial sub headings does NOT have a corresponding EML entity type and
probably should not be used since theres no way to describe its logical
structure.

With respect to physical formats, we could be very strict and say that if
adequate information for opening and parsing the data can NOT be encoded in
EML, then it is probably not a good format to use. That would allow us to
use delimited or fixed length files, ascii or binary, for which all
information for reconstructing the logical data from a filestream is
provided in the metadata (Physical module). A less strict policy would be
accept files or server connections that conform to a well-established format
for which a citation or link to the parsing documentation is provided in the
metadata for users that do not recognize the format by name. This is is not
supported in the current EML draft, although it is in the ASU draft for
physical.xsd. This is very convenient for supporting local lab access to
data, although it does enter some some subjectivity into deciding what is a
"well established" format and how long it might expect to remain so. Most of
us have software that can open a dBase 3 format file, but i doubt many of us
know where documentation on that format is maintained anywhere in the public
domain.

At CAP, tablular data are stored in a relational database and vectorGIS data
are stored either in SDE, Shapefiles or Coverages. Grid and image data are
in ERDAS imagine or ESRI Grid. The physical description in our EML
references these types so that local lab know what formats they are in. All
access through the web is mediated by a server application that knows how to
open these named formats. Queries for tabluar data are passed to the user as
a tab delimited ascii files, or as a zipped shapefile package, depending on
the type of data. Once we are fully functional, that file will be
accompanied by an EML document that describes these formats and not  the
storage formats.
Peter McCartney (peter.mccartney at asu.edu)
Center for Environmental Studies
Arizona State University
480-965-6791
-----Original Message-----
From: Eda Melendez-Colom [mailto:correoeda at hotmail.com]
Sent: Thursday, February 21, 2002 11:05 AM
To: dman at lternet.edu
Subject: a standard to publish data on the Web
Hi:

Is there a format standard to put our data on the web...meaning, has NSF
specified to any of the sites that the data files that we publish on our Web
sites must be for example ASCII comma-delimited or ASCII tab-delimited only?
If not what do you use at your site, if any standard?

I will appreciate your answers as soon as you can, since it is the theme of
discussion at my site now.

Thanks,

EDA


  _____

Hable con sus amigos en línea, pruebe MSN Messenger: http://messenger.msn.es
<http://go.msn.com/bql/hmtag1_etl_ES.asp>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mercury.nceas.ucsb.edu/ecoinformatics/pipermail/eml-dev/attachments/20020223/30d188e0/attachment.htm